Child welfare technology is constantly evolving: trends ebb and flow, policies and mandates change, and new tools become available seemingly every day. Beyond changing how work gets done, it’s also changing the way you talk about it—both about your challenges and how technology can be used to help solve them.
We’ve compiled a list of terms you’ve likely heard or read about related to technology, along with their definitions, how they relate to each other, and what they mean for child welfare.
Terms are listed in alphabetical order, or you can use the table below to jump to a specific word. And, if you want to dive deeper into a certain topic, most entries include links to additional resources to learn more.
Editor’s note: this post was originally published in 2018 but was recently updated to include emerging terms and current resources.
Artificial Intelligence (AI)
The ability of a machine to imitate human intelligence.
Digital assistants like Siri, Alexa, and Cortana, or cleaning “robots” like a Roomba, are a couple of examples of AI you might already be interacting with every day.
AI can be applied in child welfare to help automate processes and eliminate redundant work. AI can also help surface meaningful data that otherwise might have gone unnoticed (see dark data) and put critical information right at workers’ fingertips with no need to re-collect, re-interpret, or dig for data (see case discovery), which means they have more time to focus on clinical interactions with children and families.
Today, AI is advancing so rapidly that it’s hard to find the same definition (or definitive list of types of AI) twice. Potential use cases for AI change almost as often as the terminology does too! That’s why you’ll come across several different terms related to AI throughout this glossary, each with additional examples of its potential applications in child welfare.
Read our blog “Don’t Get Spooked by Artificial Intelligence for Child Welfare” or check out Gartner’s article “6 AI Myths Debunked” to learn about common misperceptions surrounding AI in child welfare.
(Note: no machine can ever replace a human’s ability to understand the complexities of each child welfare case. Technology can empower workers to discover elements of a case that might not have otherwise been found, but only a human can know what’s best for each child and family. This is a critical point to keep in mind as you read through any AI-related terms that follow.)
The ability to read and analyze an entire case file, like a human would, through a child welfare lens.
The case discovery module of Northwoods’ software Traverse® uses natural language processing to uncover dark data (see below) and automatically extract major life events, people mentions and connections, and critical concepts mentioned in case content (for example, drugs, risk factors, and protective factors). It presents a complete picture of the child or family’s past and present to safeguard their future.
We’ve published several resources to further explain case discovery and its impact on child welfare agencies:
- Infographic: Artificial Intelligence in Social Work Explained
- Blog: What Would You Ask if Your Child Welfare Case File Could Talk?
- Video: Child Welfare Software: Finding a Forever Home
- Article: The Key to Kinship: Technology Helps Keep Kids Close to Home
Since case discovery is a module of our software, you may not hear the term often outside of your conversations with Northwoods; however, you may hear about other similar ideas, like natural language processing, machine learning, or artificial intelligence (see above).
Gartner’s IT glossary defines dark data as “the information assets organizations collect, process, and store during regular business activities, but generally fail to use for other purposes. Like dark matter in physics, dark data often comprises most organizations’ universe of information assets. Thus, organizations often retain dark data for compliance purposes only.”
In the context of child welfare, dark data is typically defined as information collected and compiled from numerous sources over a long period of time that become hidden or virtually impossible to retrieve when making decisions to protect children and strengthen families.
Dark data is some of the most valuable information your agency has within its case files, but due to its sheer volume and complex nature, it is often the most difficult for social workers to manage and discover.
If the case files are not electronic, accessibility becomes an additional issue—for example, information can easily get lost in a paper case file or on sticky notes, hand-written notes, and other places workers might jot it down. Plus, the large paper files that contain this dark data are often not even in the same physical location as the worker making critical, time-sensitive decisions.
We’ve heard from several agencies, “well, we don’t have any dark data, so that’s not a problem for us!” But, as we continue talking and exploring the concept, they realize there’s a lot more hidden information than anyone ever thought (think psychological reports, court records, emails, medical reports, case notes, etc.).
We’ve created a few resources and visuals to help you better understand what dark data is, how it gets created, and why it’s so important:
- What Causes Dark Data in Child Welfare? (Infographic and blog)
- Shining Light on Dark Data in Child Welfare (eBook)
- Dark Data is Hiding in Your Child Welfare Case Files (Blog)
The automated sharing of data or files between two systems or organizations.
In child welfare, you’ll often hear this term described as “bi-directional data exchange,” as that’s how it was named in Comprehensive Child Welfare Information System (CCWIS) regulations.
Think about when you sync your banking account with a budgeting or financial app that can also help manage your money—that’s an everyday example of the type of exchange we’re describing.
Here’s another example using child welfare information systems: Our software Traverse autofills case, client, and service provider data provided by CCWIS into state and county electronic forms. Social workers can also complete additional information as needed to be made available to CCWIS. Because of the data exchange, workers spend less time filling out basic information (names, addresses, dates, etc.) on forms and more time engaging families, which increases the potential for a positive case outcome. Reducing these administrative obstacles can also help minimize social worker burnout and turnover.
Despite these benefits, sharing data can feel scary for a system that has always had privacy top of mind for such sensitive information. However, the Administration of Children & Families (ACF) encourages data sharing where appropriate for the benefits of children and families.
The Capacity Building Center for States, part of ACF’s Children’s Bureau, further explores the impact of data sharing in their article “Inventory of Innovations: Data and Child Welfare.” You can also view Casey Family Programs’ article “How Can Data Sharing Across Child- And Family-Serving Systems Be Implemented Effectively?” for practical advice to get started.
If you’ve ever uploaded a photo to Facebook and been prompted to tag someone in it, you’ve seen deep learning in action. Facebook has analyzed hundreds or thousands of photos of that person to determine what set of features make up his or her face, so it can now identify that person.
Think back to middle school science class: the human brain functions through a network of neurons, or nerve cells, that interact with each other to communicate and process information. This information determines how we operate and make decisions.
At its simplest, deep learning functions the same way: neural networks send signals to each other that help a machine process and understand very large amounts information. The more times this happens (referred to as “layers”), the more complex a conclusion the machine can make. As a result, it can teach itself to identify the features (think shapes, colors, patterns, or textures) that make up images and objects that humans recognize.
There’s an excellent article on Medium called “Neural Networks: Is Your Brain Like a Computer?” that goes into detail on this comparison. Our partners at Amazon Web Services (AWS) also have a comprehensive page “What is Deep Learning?” that answers common questions on the topic.
Here’s one way deep learning could be applied in child welfare: imagine if a machine could learn to identify a specific object in a photo in a case file, plus understand the context of that object based on other things surrounding it and then make assumptions based on all of this information (e.g., a needle is on a coffee table in someone’s home, where there are also children’s toys in the background, which means this is an unsafe living environment for a child).
Examining data or content to answer the question “What happened?” or “What is happening?”
While descriptive analytics is performed manually in many industries, tools like Traverse are now available for child welfare that can analyze large sets of data in a matter of minutes. Traverse applies natural language processing and machine learning to automatically read case files like a social worker would and highlight the most important information in a case. That way, a social worker is very quickly presented with a child’s whole story and can use it to make more informed decisions to ensure safety.
Descriptive analytics and predictive analytics are sometimes used interchangeably, but they are in fact different, and it is important for a social worker to know this distinction. Descriptive analytics surfaces information within the case, presenting it to a social worker to help guide them in making case planning decisions. However, the social work makes the final decision. Predictive analytics also surfaces information, but then translates the information into a decision that the social worker did not have to make for themselves—for example, using case data to make safety decisions and predictions.
A system or process used to capture, store, retrieve, and manage documents and files.
While some agencies unfortunately lack funding, resources, or justification to move away from paper-based systems, many have realized that electronic document solutions are critical for keeping up with increasing and complex caseloads and changing work expectations (like the ability to work remotely). An electronic document management system (EDMS) is also foundational for leveraging other emerging technologies like automation and artificial intelligence.
When evaluating document management solutions through a child welfare lens, consider the following features:
- Scan, upload, and capture documents and other case content (audio, video, or photos) to the electronic case file using agency defined taxonomy.
- Provide agency staff with immediate and simultaneous access to documents from anywhere.
- Quickly find and filter case content by date or content type, or by using full text search.
- Documents can be routed within the agency to automate child welfare processes.
- Keeps an audit trail of where documents have been routed and how they’ve been managed.
View our customer story “Wilson County Supports Workers with Modern Software for Social Services” to see the benefit of these features in action for a child welfare agency. Our blog “Every Case File, One Solution: The Future of Human Services Software” also details how a modern EDMS can help child welfare workers improve collaboration and information/data exchange with other programs.
A system or process used to manage, complete, and process forms.
Like electronic documents, many child welfare agencies have recognized the importance of digitizing their forms as a first step toward increasing their efficiency, timeliness, and service delivery capacity.
Here are some features of electronic forms that can have the greatest impact on your workers and clients in child welfare:
- Allow staff to complete forms from anywhere they conduct their work, regardless of connectivity.
- Autofill client, case, and provider information to minimize duplicate work.
- Digitally and securely share forms for review or signature, both internally and externally.
- Make all form data searchable.
- Allow for form data to be utilized in reports to gain insights on a macro level.
Read our blog “A Fresh Approach to Electronic Forms in Human Services” to learn more.
A type of artificial intelligence that can generate new content, often in the form of text, images, or other media, based on the data it was trained on.
Generative AI has recently become a key topic of discussion in human services, thanks in large part to public tools like ChatGPT that anyone with an email address can access. They offer promising capabilities, from enhancing client communication to managing case details and streamlining manual tasks, but also pose unique challenges related to data privacy and trust in their outputs.
Our blog “Artificial Intelligence and ChatGPT for Human Services and Social Work: Dos and Don’ts” provides emerging use cases and best practices to help child welfare agencies use generative AI tools responsibly and effectively, plus links to 10 industry resources to learn more. AWS has another resource page “What is Generative AI?” that answers additional questions and shares common applications of the model.
An approach to designing a product or service in which the people who actually use it are placed in the center of the process. It seems like common sense that a product or service should be designed with its prospective user in mind, but there are many instances where that’s not the case—especially when it comes to child welfare technology.
Human-centered design assures that people, processes, and products are connected. It requires a deep understanding of not just how someone is going to use a tool, but also what they expect it to help them accomplish, how they need to be trained and supported, or what potential challenges they’ll face while learning it. Human-centered systems are intrinsically linked to a child welfare agency’s ability to improve outcomes and do meaningful work. If a system isn’t user-friendly, social workers and their clients just won’t use it, which means the agency won’t experience its benefits.
Child Trends has a good post explaining how human-centered design can be applied in child welfare: “Human-Centered Design Can Create More Efficient and Effective Social Service Programs.”
We’ve also created several of our own resources to explain how Northwoods approaches this concept:
- Human-Centered Design: How Northwoods Builds Solutions Around You
- Designing Technology to Enhance Engagement & Meet Users Where They Are
- 5 Must-Haves for Mobile Technology in Child Welfare
- 7 Ways to See if Your Technology is Meeting Caseworkers’ Expectations
The ability of different IT systems and software applications to communicate, exchange data, and use the information that has been exchanged.
The concept of interoperability, initially brought to the forefront for its impact on CCWIS) data exchange requirements, has sparked an industry-wide shift in thinking about how case data and information can and should be stored and shared. Here are a couple of examples of how this can positively impact child welfare:
- Reduce duplicate data entry: the less time workers have to spend copying and pasting data from one system to another, the more time they have to focus on engaging children and families.
- Access to information: when data flows freely from one system to another, workers can piece together a more complete and holistic view of each family’s story, ensuring better quality within the case.
- “No Wrong Door” approach: child welfare workers need to collaborate and exchange information with their counterparts in programs like economic assistance, child care, and child support to create holistic support plans for families.
Our blog “How CCWIS Federal Requirements Should Spark Systemic Change” shares action steps for agencies looking to capitalize on the possibilities of interoperable systems and meaningful data exchange.
A form of artificial intelligence that teaches machines to act in specific ways without explicitly programming them to do so.
If you’ve ever ordered something on Amazon and then received a follow-up email on additional items you might like, you’ve seen machine learning in action. Amazon is using a machine-learning model to analyze what you’ve browsed, what you’ve bought, and what other people who have similar interests to you have browsed and bought, and then build recommendations based off that data.
Here’s how it works: at its core, machine learning is all about gaining insight from large amounts of data. Let’s say you have a spreadsheet with 10,000 rows of data that you split into two groups: training data and test data. A machine can learn a series of algorithms that teach it how to read and find patterns in the training data, and then use those patterns to make predictions or assumptions about the remaining test data.
Case discovery is an example of how machine learning can provide guidance to help workers understand all the information that exists, so they can apply it toward decision-making. When we talk about how Traverse can read a case file “like a social worker would,” we’re talking about machine learning: case discovery relies on machine learning to read, analyze, and extract the key concepts, events, and connections in a case.
Natural Language Processing (NLP)
A form of artificial intelligence that allows computers to understand human language.
In child welfare, using NLP to read and analyze case content can surface incredible amounts of information.
Traverse, for example, can read the entire case file—everything from documents and case notes to photos, audio, and video files. However, to make the technology even more impactful, Traverse uses NLP in conjunction with machine learning (see above) to not only read the case, but also understand it the same way a social worker would because the models have been trained specifically for child welfare. Take the phrase “Bobby beat Jenny at the race” as an example. A machine that hasn’t been trained through a child welfare lens may interpret that Bobby is very fast and won the race. However, from a child welfare perspective, this same phrase could indicate physical abuse.
Coursera’s article “What is Natural Language Processing? Definitions and Examples” breaks down how NLP works and the benefits it provides. AWS has a helpful resource page as well: “What is NLP?”
Examining data or content to predict “What is going to happen?” or “What is likely to happen?”
Predictive analytics use machine learning to generate a predictive model, allowing people to discover patterns and trends based on historical data, say living within a case file.
It’s been said that predictive analytics aims to improve child welfare outcomes, but this model has also raised various concerns, including bias. For example, predictive analytics tools once promised that they could help agencies better understand which groups of children in their communities may have an increased chance of being in danger of abuse. This may be true, but practitioners quickly realized the risks of implementing this sort of technology could outweigh the reward.
Youth Today’s article “Predictive Analytics in Child Welfare Raise Concerns” discusses how predictive analytics tools can perpetuate and amplify biases that harm communities of color. The McSilver Institute for Poverty Policy and Research at New York University further debates the topic in their article “Experts Discuss Pros and Cons of Predictive Risk Tools in Child Welfare Practice.”
It’s worth reiterating here that regardless of how much we advance our work with data and technology, machines alone can’t answer the entire need. Predictive analytics tools can provide insight, but social workers must be able to apply their own training, observation, and critical thinking skills to understand how the data applies to each unique situation and case.
It’s also important to remember here that context is critically important. The predictions and precautions brought forward by this type of technology are only as good as the data the tool has analyzed. If the data contains bias, the predictions will reflect that bias.
Robotic Process Automation (RPA)
A form of automation that uses software bots or artificial intelligence agents to automate repetitive tasks.
RPA emerged ahead of the COVID-19 pandemic as one of the next-generation technologies with significant potential to transform how government agencies work. Fast forward to today and this potential is even greater. Using RPA to streamline processes and automate routine tasks, caseworkers can have more time to focus on engaging clients and doing mission-critical work.
While most conversations surrounding RPA focus on means-tested eligibility programs, our blog “Putting the ‘Human’ Back in Human Services Through Robotic Process Automation” shares some potential use cases for child welfare, such as:
- Supporting compliance by simplifying audit processes—for example, a bot could find cases that meet review requirements or help verify that proper documents/evidence are attached to files.
- Moving cases forward faster by ensuring that all necessary documentation is in place prior to transferring, closing, or filing a case in court.
- Facilitating cross-program information and data exchange by identifying changes made in one system, such as integrated eligibility, and applying them to another, such as child welfare.
Software as a Service (SaaS)
Gartner’s IT glossary defines SaaS as “software that is owned, delivered, and managed remotely by one or more providers.” You may also hear of these tools referred to as web-based or hosted software.
Salesforce’s resource page “What is SaaS?” further clarifies that “Instead of installing and maintaining software, you simply access it via the Internet, freeing yourself from complex software and hardware management.”
You can see the benefit of this model for child welfare agencies with limited IT resources, as you essentially rent both your software licenses and the servers/infrastructure needed to host the solution. This means the technology vendor that provides the solution will cover things like security, support, and storage. SaaS solutions are also built to continuously evolve, which can be especially helpful for child welfare agencies trying to keep up with ever-changing policies and requirements or support a hybrid workforce.
AWS’s resource page “What is SaaS?” goes into more technical detail on how the model works and is typically applied. Our blog “Answering Your Commonly Asked Cloud Questions in Human Services” also adds more context through a child welfare lens. We also address security-related concerns that many agencies associate with SaaS software in our post on “Solving Cybersecurity Problems in Human Services Before They Start.”
Structured DataInformation or content that has a pre-defined data model and is organized in a pre-defined manner.
For child welfare, because structured data fits neatly into a database and is easily searchable (think names, dates of birth, or demographic information on a form field), it is what gets added to the system of record.
Northwoods' customers have confirmed that only 20% of case-related data is structured, which means most of the information can be dark data or unstructured data that lives elsewhere and can be more difficult to find and use when making decisions.
System of Engagement
Technology systems designed to meet the business needs of human services agencies and provide workers with the tools they need to produce more efficient and effective client-centered outcomes.
A system of engagement features an intuitive, human-centered design that supports collaboration, enhances daily work, and enables workers to spend as much time as possible delivering services. It works together with an agency’s existing systems, such as CCWIS, to enhance how workers manage information and interact with children and families, regardless of their location or level of connectivity.
A purpose-built system of engagement eliminates redundant work, allowing workers to repurpose time previously spent on administrative tasks to do more high-value work with children and families. It also helps them uncover insight (see dark data and descriptive analytics) that can be applied when making critical decisions.
Read “Enhancing CCWIS with a System of Engagement” or “5 Ways a System of Engagement Benefits Child Welfare” to learn more.
System of Record
An information storage system that is the authoritative data source for a set of data elements.
For child welfare, a system of record is built to store data needed to generate reports for the federal government. It is not necessarily intended to help a social worker work through the case. Even when workers put information into a system of record, it’s a synopsis of what happened, not the actual documentation of the event. It summarizes the data; it isn’t the data itself.
When it comes to CCWIS, we believe states should incorporate both a system of record and a system of engagement to maximize efficiency and effectiveness.
Since many documents, and therefore critical information, are often stored in other systems, augmenting your system of record with a system of engagement (see above) can help your agency maximize efficiency and effectiveness. Our video “How Does Traverse Support State Child Welfare Case Management Systems?” explains what this idea looks like when put into action.
Classification or organization of terms and phrases that represents a specific domain of information.
As it relates to child welfare and document management, taxonomy is just a fancy word to describe how electronic files are named and organized in a case file (you might compare this to the tabs along the edge of a paper case file to help you locate a specific document quickly). It’s by no means a new term, but it remains relevant as technology changes. Having a well-structured taxonomy is foundational to help workers locate the right information at the right time and in the right context.
Think about it this way: all data needs to be organized and accessible in a way that makes sense to child welfare workers and supervisors when they’re faced with a decision. A fire hose of jumbled information isn’t helpful at all. Instead, use taxonomy to guide workers down the path to the right information that’s most important to the case.
Read “How a Well Designed Taxonomy Can Increase Your Agency’s ECM Success” to learn more.
Information or content that either does not have a pre-defined data model or is not organized in a pre-defined manner.
Think of unstructured data as providing the context, or the story, around the structured data, so it’s very relevant when making decisions. However, it doesn’t fit neatly into a database and it isn’t always easy for a machine to analyze, so it can easily become dark data.
Northwoods' customers have confirmed that 80% of the information and data that completes a case file is in an unstructured format (think photos, emails, psychological reports, medical reports, releases of information, home assessments, etc.).
When you consider this, tools that can turn unstructured data into insight, like a system of engagement or artificial intelligence, become even more useful because they can identify and extract additional data that wouldn’t have been entered into a system of record.